Tetrapod Phylogeny Inferred from 18S and 28S Ribosomal RNA Sequences and a Review of the Evidence for Amniote Relationships1 S
Total Page:16
File Type:pdf, Size:1020Kb
Tetrapod Phylogeny Inferred from 18s and 28s Ribosomal RNA Sequences and a Review of the Evidence for Amniote Relationships1 S. Blair Hedgest * Kirk D. Moberg,? and Linda R. Maxson* *Department of Biology and Institute of Molecular Evolutionary Genetics, Pennsylvania State University, and ?University of Illinois College of Medicine at Urbana-Champaign The 18s ribosomal RNAs of 21 tetrapods were sequenced and aligned with five published tetrapod sequences. When the coelacanth was used as an outgroup, Lis- samphibia (living amphibians) and Amniota (amniotes) were found to be statis- tically significant monophyletic groups. Although little resolution was obtained among the lissamphibian taxa, the amniote sequences support a sister-group rela- tionship between birds and mammals. Portions of the 28s ribosomal RNA (rRNA) molecule in 11 tetrapods also were sequenced, although the phylogenetic results were inconclusive. In contrast to previous studies, deletion or down-weighting of base-paired sites were found to have little effect on phylogenetic relationships. Mo- lecular evidence for amniote relationships is reviewed, showing that three genes (beta-hemoglobin, myoglobin, and 18s rRNA) unambiguously support a bird- mammal relationship, compared with one gene (histone H2B) that favors a bird- crocodilian clade. Separate analyses of four other genes (alphacrystallin A, alpha- hemoglobin, insulin, and 28s rRNA) and a combined analysis of all sequence data are inconclusive, in that different groups are defined in different analyses and none are strongly supported. It is suggested that until sequences become available from a broader array of taxa, the molecular evidence is best evaluated at the level of individual genes, with emphasis placed on those studies with the greatest number of taxa and sites. When this is done, a bird-mammal relationship is most strongly supported. When regarded in combination with the morphological evidence for this association, it must be considered at least as plausible as a bird-crocodilian relationship. Introduction The classical view of tetrapod relationships (fig. 1) is based largely on fossil evidence (Romer 1966; Carroll 1988; Gauthier et al. 1988). However, morphological and molecular data from living species recently have challenged some long-held beliefs. The most controversial suggestioh is that mammals, not crocodilians, may be the clqsest relatives of birds (Gardiner 1982; Lravtrup 1985 ) . Moreover, the single origin of the Lissamphibia (frogs, salamanders, and caecilians) never has been firmly estab- lished (Parsons and Williams 1963; Jarvik 1968, 1980; Lravtrup 1985; Milner 1988; Panchen and Smithson 1988). To address these questions concerning tetrapod phy- logeny, we have examined the phylogenetic relationships of 26 tetrapods representing 1. Key words: Lissamphibia, Amniota, Reptilia, Aves, Mammalia, crystallin, hemoglobin, histone, immunoglobin, insulin, myoglobin, molecular. This article is contribution 1 from the Institute of Molecular Evolutionary Genetics at Pennsylvania State University. Address for correspondence and reprints: Dr. S. Blair Hedges, Department of Biology, 208 Mueller Lab, Pennsylvania State University, University Park, Pennsylvania 16802. Mol. Biol. Evol. 7(6):607-633. 1990. Q 1990 by The University of Chicago. All rights reserved. 0737-4038/90/0706MX)7SO2.00 608 Hedges et al. .---- turtles . _-- sphenodontlda - .---- ----------- squsmatea smn\~!e?::- -:,..=- - crocodilians , -- ---- ----------------- birds ---- ----____ mammala - - - - -- - - - - - -- - -- -- --caecilian~ salamanders ----___ frogs Million years before present FIG. 1.--Currently recognized relationships of the major tetrapod groups, based largely on the fossil record (Romer 1966; Carroll 1988). Solid lines indicate continuous fossil record; dashed lines denote either the absence of fossils (of the major groups) or inferred transitions. nearly all major lineages, by using nucleotide sequences of the slow-evolving 18s ribosomal RNA (rRNA) molecules. Additional sequence data also were obtained from portions of the 28s rRNA subunit in 11 tetrapods. Methods We obtained sequences from 15 species of amphibians representing 14 families (seven frogs, four salamanders, and four caecilians), a turtle, a crocodilian, two squa- mates, and two birds by direct sequencing of nuclear 18s rRNA. The anuran taxa are Bufonidae (Bufo valliceps), Discoglossidae (Discoglossus pictus) , Hylidae (Hyla ci- nerea), Leptodactylidae (Elactherodactylus cuneatus), Microhylidae (Gastrophryne carolinensis), Pelobatidae (Scaphiopus holbrooki), and Sooglossidae (Nesomantis thomasseti) . The caecilians are Caeciliaidae-1 ( Grandisonia alternans) , Caeciliaidae- 2 (Hypogeophis rostratus), Ichthyophiidae (Ichthyophis bannanicus), and Typhlo- nectidae ( Typhlonectes natans). The salamanders are Ambystomatidae (Ambystoma mexicanum), Amphiumidae (Amphiuma tridactylum), Plethodontidae (Plethodon yonahlossee), and Sirenidae (Siren intermedia). The amniotes are a turtle (Pseudemys scripta ) , an alligator (Alligator mississippiensis), a lizard ( Sceloporus undulatus) , a snake (Heterodon platyrhinos), and two birds [Galliformes (Gallus) and Passeriformes (Turdus migratorius)]. These were compared with published sequences of a pipid frog (Xenopus laevis; Salim and Mayden 198 1) and of four mammals: Oryctolagus cuniculus (Rairkar et al. 1988), Rattus nowegicus (Torczynski et al. 1983; Chan et al. 1984), Mus musculus (Raynal et al. 1984), and Homo sapiens (Torczynski et al. 1985). This broad sampling of taxonomic diversity within each of the three orders of amphibians permits the distinction between those sites that are unique to a single lineage and those more informative sites that are representative of an entire group. The sequence from the only living actinistian fish, the coelacanth (Stock et al., ac- cepted), was used to provide an outgroup. Some of the same taxa (and samples) used to obtain 18s rRNA sequences also were sequenced at the 28s rRNA gene. They include Bu fo, Discoglossus, Hyla, Typhlonectes, Ambystoma, Pseudemys, Sceloporus, Heterodon, Alligator, Turdus, and Gallus. Published 28s rRNA sequences of Xenopus (Ware et al. 1983), Mus (Hassouna et al. 1984), and Homo (Gonzalez et al. 1985) were used for comparison. Sequenced were three regions of the 28s rRNA correspond- ing to the following sites in Mus (Hassouna et al. 1984): 1-355, 2157-26 1 1, and 4277-46 1 1. rRNA was isolated from -- lg of liver (or, if available, lg of ova). The tissue was homogenized in a buffer [lo mM Tris hydroxymethyl amino methane-HC1 pH 8.0, Tetrapod Phylogeny and Amniote Relationships 609 1 mM ethylene diamine tetra-acetate pH 8.0, 2% (w/v) sodium dodecyl sulfate, 5% ( W/V)sodium tri-isopropylnaphthalene-sulfonate] by using a Brinkman Polytron. The homogenate was extracted twice with phenol, once with phenol/chloroform, and once with chloroform, followed by ethanol precipitation and resuspension of the pellet in diethylpyrocarbonate (DEW)-treated water. Direct sequencing of the extracted rRNA (by using reverse transcriptase) followed the procedure of Lane et al. ( 1985), although the chase step was not found to be useful and thus was omitted. A combination of 17 oligonucleotide primers allowed us to generate (a) nearly unintempted sequences from the entire 18s rRNA subunit ( -- 1,900 bp) in each of the species and (b) -400 bp from portions of the 28s molecule. The 18s primers used and their starting positions in Mus (Raynal et al. 1984) are CTAGAATT( AG)CCACAGTTATCC ( 145), TACCATCGAAAGTTGATAGGC- AGA ( 354), ACCGGCGGCTGCTGGC (614), GTCCTATTCCATTATTCC (860), CCG( AG) TCCAAGAATTTCACCTCT (956 ), GCCCTTCCGTCAATTCCTTTA- AGTTTCAGC ( 11 85 ) , GTCAAATTAAGCCGC ( 123 3 ) , AAGAACGGCCAT- GCACCACC ( 1324), TCTAAGGGCATCACAGACCTGTTATTG ( 1482), ACG- GGCGGTGTGAC ( 1693), and GGTTACCTTGTTACGACTT ( 1824). The 28s primers and their starting position in Mus (Hassouna et al. 1984) are GTTGGTTTCTmCCT ( 56), TTTGGGCTGCATTCCA (29 1 ) , CTTTCCC- TCACGGTA ( 365 ) , CTTGGAGACCTGCTGCGG (2543 ) , CCTTATCCCGAAG- TTACG (26 19 ) , and CAGGTCGTCTACGAATG (46 13). Alignments initially were done with Intelligenetics software, although refinement of the entire alignment was made by eye, in order to increase sequence similarity. ' Regions that could not be aligned, because of length or sequence variability, were omitted from the analyses. Ambiguities (multiple bands at the same site on a gel) were scored as "N and were treated in the parsimony analyses as missing data and were not used in the calculation of pairwise distances for the neighbor-joining (NJ) analyses. The informative sites (those with at least two bases, each occurring in more than one taxon) were analyzed with the maximum-parsimony (MP) method in PAUP (Phylogenetic Analysis Using Parsimony, version 3.0). Values for consistency index (CI) presented do not include sites with unique variants. The NJ method (Saitou and Nei 1987), version 2.0, was used to generate trees from distance matrices (corrected for multiple substitutions). The statistical significance of groups in both the MP and NJ analyses was evaluated by the bootstrap method (Felsenstein 1985), with 1,000 iterations. Amino acid sequences of seven genes relevant to amniote relationships were obtained from the National Biomedical Research Foundation Protein Identification Resource (PIR) data bank (version 21 ), from the Swiss Protein data bank (version 1 1), and from Lance et al. ( 1984). The data-base locus names for the sequences used in the combined analysis (see